PDA

View Full Version : Consolidated NV40 vs. R420 benchmark results


Pages : [1] 2 3 4 5 6 7 8 9

MikeC
05-04-04, 07:59 PM
I spent most of the evening putting together this chart by comparing the benchmark results in each review listed at the top of the chart. The GeForce 6800 Ultra results are for the non-overclocked Ultra model.

I wanted to make sure the information was reliable and I double checked the results. Let me know if there is anything that needs to be changed. Too tired to make any conclusions yet :)

http://www.nvnews.net/images/news/200405/comparison.png


Update - 5/5/2004: Included results from X-bit labs' preview. Incorporated various cosmetic changes to chart layout.

Update - 5/8/2004: Updated the chart based on feedback from a discussion thread over at Beyond3D:

http://www.beyond3d.com/forum/viewtopic.php?t=12186

NVIDIA changed to ATI for UT2004 at Hot Hardware. NVIDIA changed to DRAW for Splinter Cell at FiringSquad.

Update - 5/8/2004: Created a new spreadsheet with more details.

http://www.nvnews.net/vbulletin/showthread.php?t=28520

eesa
05-04-04, 08:00 PM
wow I'm seriously impressed.
:clap2: :clap2:

Stevia
05-04-04, 08:03 PM
nice work Mike :jumping:

MikeC
05-04-04, 08:03 PM
wow I'm seriously impressed.
:clap2: :clap2:

Thanks :) Some of the websites results were too hard to decipher.

I only included results from reviews that did head-to-head comparisons using the same settings.

There are a couple that may be misleading, but I didn't have the time to fix them right now. For example, Tom's Hardware shows an NVIDIA win in Halo only becuase they didn't benchmark with AF.

SH64
05-04-04, 08:03 PM
@Hardware Analysis .. the 6800U is winning in FC
http://www.hardwareanalysis.com/content/article/1710.8/

MikeC
05-04-04, 08:06 PM
@Hardware Analysis .. the 6800U is winning in FC
http://www.hardwareanalysis.com/content/article/1710.8/

I didn't include their results since the 6800U core was clocked at 450MHz.

"This GeForce 6800 Ultra sample board is different only by the clockspeed it runs at, 450-MHz instead of 400-MHz."

SH64
05-04-04, 08:07 PM
Oh & very nice work Mike :super: .. you deserve a medal for that .. errr or a 6800U-EXT if you like :D

EDIT : oops sorry didnt notice that .

eesa
05-04-04, 08:10 PM
Thanks :) Some of the websites results were too hard to decipher.

yeah, I would imagine not exactly everything neatly fits into a category.

flakjacket
05-04-04, 08:21 PM
What resolution, AA/AF etc was considered? It seems that it goes back and forth between settings in a lot of reviews.

SH64
05-04-04, 08:22 PM
Total results : Nvidia = 26 , ATI = 24 , Draw = 8 .

XanderF
05-04-04, 08:23 PM
It would be interesting to see a sub-chart showing with cards win which API.

IE., how many of the nVidia wins were DX vs OpenGL, how many of the ATI wins were DX vs OpenGL.

eesa
05-04-04, 08:24 PM
Total results : Nvidia = 26 , ATI = 24 , Draw = 8 .

that's doesn't mean much though. Which of those do you personally actually care about?

eesa
05-04-04, 08:25 PM
btw, mike, how did you decide who won or lost in situations where AA/AF or even resolution turned the tables?

SH64
05-04-04, 08:26 PM
that's doesn't mean much though. Which of those do you personally actually care about?

Its not important what I care about .. because every gamer has his own interest .
so this give a general result for "games that are in the market" on 2 diffrenet video cards .
end of story .

jAkUp
05-04-04, 08:27 PM
wow.. thats nice. good job mike

MikeC
05-04-04, 08:29 PM
What resolution, AA/AF etc was considered? It seems that it goes back and forth between settings in a lot of reviews.

I summarized the results across all resolutions and AA/AF settings. Most web sites present their results with bar charts that are ordered from fastest to slowest. My conclusion was based on a casual examination of the charts in each review.

For example, if one card ran faster without AA/AF and the other ran faster with AA/AF, I may have called it a draw. I didn't want to make this comparison too complicated. Feel free to draw your own conclusions, but tell me if you think I am way off the mark on a particular comparison.

flakjacket
05-04-04, 08:31 PM
I thought the extremetech review was a little different from the others results. But good job! How did you determine a winner?

flakjacket
05-04-04, 08:34 PM
It's also weird that only Tom's used Colin Rally 4, doesn't that game use dx9? I wonder why it wasn't widely used?

SH64
05-04-04, 08:36 PM
It's also weird that only Tom's used Colin Rally 4, doesn't that game use dx9? I wonder why it wasn't widely used?

No CMR 04 is basicly a DX8 game . it uses PS1.1 if i'm not mistaken .

MikeC
05-04-04, 08:38 PM
I thought the extremetech review was a little different from the others results. But good job! How did you determine a winner?

I've always felt that performance boils down to the game or type of game being tested. So my answer would be that it depends :)

But there are many other factors other than performance to consider.

-=DVS=-
05-04-04, 09:51 PM
Nice job Mike , to bad review sites only bench moust common games , benching some uncommon games would be interesting also where is my Lineage 2 benchmark :(

ragejg
05-04-04, 09:55 PM
damn Mike, nice job!! That chart has many uses!! :firedevil

btw: @ DVS: yeah, wth!! Where's ninja turtles, Silent hill 3, XIII, Savage... :p

Spotch
05-04-04, 10:15 PM
I didn't include their results since the 6800U core was clocked at 450MHz.

"This GeForce 6800 Ultra sample board is different only by the clockspeed it runs at, 450-MHz instead of 400-MHz."

That guy is on crack. Unless the 6800 really does perform better at 1600x1200 than at 1024x768 in UT2K4. :angel2:

Snarfy
05-04-04, 10:21 PM
why arent the hardocp results there?

besides that, awesome chart! :)

eesa
05-04-04, 10:22 PM
why arent the hardocp results there?


errr... runs away.

be gentle...